How we protect conversations

WHISPAR watches for signs of real distress in three steps. A few phrases trigger an instant, local check as you type — before anything even leaves your device. Separately, a behavioral monitor watches the shape of a conversation (pacing, one-sidedness) without reading its content. If that raises a concern, an AI model reads the recent messages in that moment to check for genuine crisis signs and can offer support. Only a score is ever kept afterward — never the words themselves.

Crisis resources

If you're going through something serious, we want you to have real support. findahelpline.com connects you to crisis resources in your country. In the US: text HOME to 741741.

Reporting

Every conversation has a report button. Use it — a report factors into our automated safety and ban systems immediately. We're still building a dedicated human review queue for individual reports; today they inform our systems in aggregate rather than being read one by one. Repeated bad behavior (like sharing contact info to move someone off-platform) is met with automatic, escalating bans regardless of whether it was individually reported.

Age requirement

WHISPAR is for adults 18 and older. New arrivals also pass an invisible, automated check before they can start a conversation, to keep bots and scripted abuse out — real visitors never see it or have to do anything.

Who to contact

safety@xeres.tech